machine design, Vol.4(2012) No.4, ISSN 1821-1259 pp. 231-236

machine design, Vol.4(2012) No.4, ISSN 1821-1259
pp. 231-236
Research paper
METHODOLOGY OF DETERMINATION THE INFLUENCE OF CORROSION PIT ON
DECREASE OF HYDRO TURBINE SHAFT FATIGUE LIFE
Dejan MOMČILOVIĆ1, * - Radivoje MITROVIĆ2 - Ivana ATANASOVSKA3 - Tomaž VUHERER4
Institute IMS, Bulevar vojvode Mišića 43, 11000 Belgrade, Serbia
University of Belgrade, Faculty of Mechanical Engineering, Kraljice Marije 16, 11120 Belgrade, Serbia
3 Institut Kirilo Savic, Vojvode Stepe 51, 11000 Belgrade, Serbia
4 University of Maribor, Faculty of Mechanical Engineering, Smetanova Ulica 17, 2000 Maribor, Slovenia
1
2
Received (25.05.2012); Revised (18.11.2012); Accepted (23.11.2012)
Abstract: This paper describes the influence of corrosion on stress concentration factor and crack initiation at shaftflange transition section. The case study of hydraulic turbine shaft failure is used as the basis for this research.
The quantification of the stress concentrators was accomplished by the usage of Theory of critical distances (TCD) in
the prediction of high-cycle fatigue behavior in machine parts and systems. The stresses obtained by Finite Element
Analysis, was used as an entry values for application of Theory of critical distances. The TCD represents a major
extension of linear elastic fracture mechanics, allowing it to be used for short cracks as well as for stress
concentrations of arbitrary geometry. Presented methodology is particular valid for machine parts of non-standard
dimensions. The significance of results presented in this paper is that TCD and developed methodology can be used in
preventing failures in power industry.
Key words: fatigue, fracture, turbine shaft, Theory of Critical Distances
1. INTRODUCTION
The most frequent causes of failure are defects such as
pores or cracks introduced during manufacturing and
inadequate design features which cause excessive stress
concentration, such as sharp corners and other notch like
type of stress raisers. In all such cases the common factor
is high local stresses and also stress gradients, the stress
decreasing with distance from the feature. One of the key
points in failure analysis is the ability to make accurate
predictions of the strength and fracture of materials in
complex load-bearing structures. This task, the strength
calculation of failed part, is also linked with the review of
original designer approach in order to check hidden
design errors. The majority of failures could be addressed
to high-cycle fatigue [1, 2], and due to the fact that real
engineering structures contain stress concentrations from
which fatigue cracks frequently initiate, the prediction of
the effect of stress concentrations on fatigue life and
fatigue strength is of great importance.
Despite this, there is no commonly accepted set of
standard methods for predicting the effect of notches,
holes, joints, defects and other stress-raisers. This is not
due to a shortage of theories and methodologies – there
are, in fact, a multiplicity of methods to be found in the
literature for the prediction of notch fatigue behavior. It is
rather that the scientific community has not been able to
decide which method, or methods, are most suitable, and
under which circumstances [1, 3, 4]. The common point
in all that theories the assumption that the relevant
parameter is the stress averaged over some critical
volume, but for convenience of calculation this has
usually been simplified to the stress at one point (a given
distance from the hot spot) or to the stress averaged over a
line of given length. The origins of above stated are in
work of Peterson and Neuber. Peterson considered that
fatigue failure occurs when the stress at a constant
distance beneath the surface equals the fatigue strength of
the material [5] while Neuber related the stress
concentration factor at the root of a sharp notch to the
average stress over a certain critical distance [6].
Summary of the materials response under fatigue is
presented by Suresh [7] where is clearly demonstrated
that process of fatigue starts with the first cycle. The local
regions of high stress promote the basic process of crack
nucleation during high cycle fatigue. The damage
progresses through mechanisms starting with crack
nucleation, and the growth of micro structurally small
cracks. Each mechanism is associated with a
characteristic size or scale of magnitude, and each
characteristic size has its own geometric complexity,
constitutive law, and heterogeneity. Fatigue behavior
cannot be fully understood and predicted without
obtaining information about each of the characteristic
sizes. According to the observations of Kitagawa and
Takahashi [3], fatigue crack growth behavior of short
cracks differs in a non-conservative manner from
expectations based upon long crack behavior. Further
contribution in understanding was done by El Haddad [8]
who introduced an effective crack length to predict the
propagation behavior of short cracks. In this formulation
the effective crack length is equal to the crack length plus
an amount “l0”, which is related to the critical distance.
This value “l0” was stated by the authors as a
characteristic of the material and the material condition
accounting for the non-continuum behavior of very small
cracks.
*Correspondence Author’s Address: Institute IMS, Bulevar vojvode Mišića 43, 11000 Belgrade, Serbia,
[email protected]
Dejan Momčilović, Radivoje Mitrović, Ivana Atanasovska, Tomaž Vuherer: Methodology of Determination the Influence of Corrosion Pit
on Decrease of Hydro Turbine Shaft Fatigue Life; Machine Design, Vol.4(2012) No.4, ISSN 1821-1259; pp. 231-236
The theory of critical distances (TCD) as popularized by
Taylor [9] attempts to predict the effect of notches and
other stress concentration features by considering the
stress field in the region close to the notch tip. This theory
requires two parameters, a characteristic distance and a
critical stress or strain characterizing failure. In one
version of the critical distance theory, termed as the Point
Method, the failure occurs when the stress becomes equal
to the failure stress at a given distance from the notch
root. In the other version of the critical distance theory,
termed as the Line Method, the failure is assumed to
occur when the stress becomes equal to the failure stress
when computed as an average value over a line of given
length. The background philosophy is lying behind the
TCD is described as wish to observe engineering
components rather than to test specimens. In practice this
meant that we only considered predictive methods which
could be applied to bodies of arbitrary shape and size,
subjected to arbitrary loadings, containing stress
concentration features of arbitrary geometry. This is
achieved by measuring material behavior using test
specimens containing notches rather than cracks (fatigue
threshold ∆Kth and toughness Kc using sharp notches
rather than pre-cracks) which avoids the difficulties and
uncertainties of carrying out standard fracture mechanics
tests. The second presumption of successful application of
TCD is the existence of an accurate stress analysis of the
machine part.
Methodology of application of TCD is as follows:
1. Generating stress–distance diagram such as fig.1 and
read the necessary stress values from FEA; and
2. The second step is calculating the value of the critical
distance.
2. EXPERIMENTAL PROCEDURE
The problem of quantifying the influence of stress raiser
is the problem that could be used in failure analysis. In
order to check the usability of TCD in real case study
instead of laboratory specimens, we chose to re-examine
failure of hydro turbine shaft [11].
Briefly, failure analysis of hydro-turbine shaft, fig 2,
shows that there is a zone from shaft-to-flange transition
radius i.e. the width of the maximum stress zone, obtained
by the Finite element analysis, strongly correlates with the
width of the zone determined by non-destructive
inspection, fig 3.
CRACKING
ZONE
a)
Fig.1. Schematic explanation of TCD point method
The result is that the critical distance for the point method
is L/2, [1, 3, 4, 9], where:
L=
1 ⎛ ∆K th
⎜
π ⎜⎝ ∆σ 0
⎞
⎟⎟
⎠
2
(1)
In this equation ∆σ0 is the fatigue limit of standard,
unnotched specimens of the material, and ∆Kth is
threshold stress intensity. The fact is that is often difficult
to define the accurate fatigue limit, the stress range
corresponding to a given number of cycles. The range
from 1 to 10 million is generally used for determination of
fatigue limit. It has been demonstrated by comparison
with experimental data that the use of this value of ∆σ0 for
L gives good predictions in many different materials, [9].
The another good example for application of TCD is
presented by Chattopadhyay [10]
232
b)
Fig.2. a) turbine shaft with runner during assembling in
the bulb, b) fractured surface
Dejan Momčilović, Radivoje Mitrović, Ivana Atanasovska, Tomaž Vuherer: Methodology of Determination the Influence of Corrosion Pit
on Decrease of Hydro Turbine Shaft Fatigue Life; Machine Design, Vol.4(2012) No.4, ISSN 1821-1259; pp. 231-236
Fig.3. Corrosion cracks at shaft to flange radius
The results of basic mechanical properties for turbine
shaft steel casting is given at the Table 1 and the results
for fatigue properties of 20GSL is given at Table 2.
3. DISCUSSION AND CONCLUSIONS
Presented case study is another confirmation, the
applicability range of TCD as a method that combines
properties of material and geometrical features of
engineering component is very wide. Compared with
standard models, which are not generally used to predict
cracking processes because they are not well adapted to
deal with stress singularities, the TCD can be used as a
complementary methodology to modern damage
mechanics. On fig. 5 the comparative stress gradients
from tension as result of axial hydraulic force during
starting regime (Fa=5542.65·103 N) without corrosion pit
and with corrosion pit are presented.
Stress (MPa)
The main conclusions of shaft failure analysis, fig. 2a),
was that the combination of several factors leads to
failure:
Inappropriate corrosion protection in the zone of
critical radius and lack of procedures of renewing
corrosion protection of turbine shaft.
Corrosion, i.e. corrosion fatigue due to leakage of
river water through the sealing box.
Table 1. Average and required mechanical properties at
room temperature of the 20GSL steel casting, [12]
Yield strength (MPa)
Tensile strength (MPa)
Elongation, in 2 in. (%)
Reduction of area (%)
Brinell hardness HB
Charpy-V notch, +20oC (J)
310
509
17.6
35.2
153
74.4
Required
(GOST 977)
- min. values
294
540
18
30
23.4
Table 2. Average and reference fatigue values of the
20GSL flange steel casting, [13]
Average
values
Fatigue limit
SFL (MPa)
168
Reference values
on air
225
in water
140
As an integral part of failure analysis, the finite element
analysis of turbine shaft was performed under various
conditions [14, 15]. The major conclusion was that stress
gradient during start up and regular regime had no
influence on crack initiation, fig. 4a) and fig. 4b). But the
simulation of corrosion pits with diameter of 600 µm, fig.
4c) during start up regime revealed that the stresses under
corrosion pit at distance L/2 was above the values of
fatigue strength of unnotched specimens tested in water.
The values of Kt (stress concentration factor) obtained by
FEM analysis was, for the case of LC1 (regime during
starting) the 3,97 and for case LC2 (regular operation
regime) 2,057. The value of Kt on corrosion pit on radius,
Fig 4c) was 5,97.
LOAD CASES:
regime during starting (LC1)
regular operation regime (LC2)
0
5
10
15
20
25
30
35
40
45
50
55
60
65
70
Distance from surface (mm)
a)
Stress (MPa)
Average
values
150
140
130
120
110
100
90
80
70
60
50
40
30
20
10
0
200
190
180
170
160
150
140
130
120
110
100
Stress gradient during starting (LC1)
90
Stress gradient during regular operating regime (LC2)
80
70
Stress gradient during starting with corrosion pit (LC1PIT)
60
50
40
30
20
0,0 0,5 1,0 1,5 2,0 2,5 3,0 3,5 4,0 4,5 5,0 5,5 6,0
Distance from surface (mm)
b)
c)
Fig.4. a) Stress gradients on r = 80 mm radius over two
regimes b) Stress gradients over two regimes, c)
Simulated corrosion pit with 600 µm diameter on
transition radius
233
Dejan Momčilović, Radivoje Mitrović, Ivana Atanasovska, Tomaž Vuherer: Methodology of Determination the Influence of Corrosion Pit
on Decrease of Hydro Turbine Shaft Fatigue Life; Machine Design, Vol.4(2012) No.4, ISSN 1821-1259; pp. 231-236
200
Stress gradient under corrosion pit
Stress (MPa)
180
which is extremely high value. Another interesting
viewpoint regarding influence of corrosion pit on
premature failure is presented on fig. 7.
160
140
120
L/2
100
0,0
0,5
1,0
1,5
2,0
2,5
3,0
3,5
Distance from the bottom of corrosion pit (mm)
Fig.5. Stress gradient under the corrosion pit
The obtained numerical results for stress state in critical
zone shows maximum Von Mises equivalent stress of
σeqv-max=230 MPa, fig. 6.
a)
b)
Fig.6. a) Meshing arround the simulated pit
b) Equivalent stresses in the zone of corrosion crack
Combined stress coefficient factor for the corrosion pit on
the radius is: Kt(sum) = Kt(radius) · Kt(pit) = 3,97 · 5,97 = 23,7
234
Fig.7. Matching of angle of fatigue fracture (throughthickness crack) and the angle obtained by FEM
simulation of corrosion pit
On the fig. 6, it is possible to see that the angle between
perpendicular axis of the shaft and the axis drawn on the
half of the corrosion pit is approximate equal to the angle
of 25 degrees. This value is almost identical to the slope
of fractured surface presented at the fig. 2.
By using the eq. (1) we obtained the value of L = 0,7 and
this value corresponds to material by available literature
data for similar cast steels and particularly with
microstructure, fig. 8 of turbine shaft material. The
microstructure of the material is the cast ferrite-pearlite
one, with oxide type non-metallic inclusions and with
minimal participation of dendrite structure. This
inhomogeneity of microstructure also explains the
formation of "pits" on transition radius.
The comparisons between stress concentration factors
and stress states obtained by Finite Element Analysis and
analytical stress calculations that used the Peterson's
elastic stress concentration factor charts for conventional
values of fillet shaft's transition radii led to conclusion
that analytical approach was not suitable for calculation
of load capacity of such turbine shaft with flange [15]. In
cases like the presented one, only FEA analysis is valid
an the use of "classic" analytical approach can lead to the
inaccurate results and wrong conclusions.
The comparisons between stress concentration factors
and stress states obtained by Finite Element Analysis and
analytical stress calculations that used the Peterson's
elastic stress concentration factor charts for conventional
values of fillet shaft's transition radii led to conclusion
that analytical approach was not suitable for calculation
of load capacity of such turbine shaft with flange [15]. In
cases like the presented one, only FEA analysis is valid
and the use of "classic" analytical approach can lead to
the inaccurate results and wrong conclusions.
The presented research points out the new trend in stress
concentration factor calculation. This procedure is
particularly suitable and important for machine parts with
non-standard geometry and specific operating conditions.
Dejan Momčilović, Radivoje Mitrović, Ivana Atanasovska, Tomaž Vuherer: Methodology of Determination the Influence of Corrosion Pit
on Decrease of Hydro Turbine Shaft Fatigue Life; Machine Design, Vol.4(2012) No.4, ISSN 1821-1259; pp. 231-236
The significance of TCD in novel, better design of
machine parts, lays in fact that the value of L is obtained
combining the material parameters such as ∆σ0 (the
fatigue limit of standard, unnotched specimens of the
material), and ∆Kth (fatigue threshold stress intensity)
with the stress gradient of particular stress raiser. The
range of applicability of this methodology is very wide.
Although the theoretical background of the TCD is
complex the use itself of TCD is simple as presented.
a)
Fig.9. Fine cracks in the critical zone of shaft-flange
transition zone, found by magnetic particle testing
after fine grinding.
b)
From above presented, it is confirmed it is again
confirmed and documented that TCD methodology has
significant potential in quantification of any type of stress
raiser influence on decrease of machine parts fatigue life.
The next step and the challenge for TCD is the
calculation the number of cycles when a crack will occur
on arbitrary geometry of machine part.
ACKNOWLEDGMENTS
The work has been funded by the Ministry of Education
and Science of Republic of Serbia, Grant TR 35029:
Development of the methodology for working capacity,
reliability and energy efficiency improvement of
mechanical systems in power industry.
c)
Fig.8. Microstructure of flange material
a). Cast, ferrite-pearlite microstructure, x 200, 5% nital
b). Micro cracks emanating from shrinkage, x 500,
c) The relative siye of microcracks
5% nital
For this particular type of turbine shafts, another
researcher gain the similar conclusions regarding crack
initiations on transition radius [16, 17, 18]. Metallurgical
approach in failure analysis, by matching appearance of
cracks, presented on fig. 9, by literature [19, 20, 21] now
is extended by calculation which is more direct proof in
cases like this one.
REFERENCES
[1] Taylor, D (2005), Analysis of fatigue failures in
components using the theory of critical distances,
Engineering Failure Analysis 12, 906–914
[2] Brooks, R,C. (1993), Metallurgical failure analysis.
New York: McGraw-Hill; pp 6-11
[3] Taylor, D (2011), Applications of the theory of
critical distances in failure analysis, Engineering
Failure Analysis 18, 543–549
[4] D.Taylor et a.l., (2011), Giornata IGF Forni di Sopra
(UD), Italy, March 7-9, 2011; ISBN 978-88-9594035-9, pp 129-135
235
Dejan Momčilović, Radivoje Mitrović, Ivana Atanasovska, Tomaž Vuherer: Methodology of Determination the Influence of Corrosion Pit
on Decrease of Hydro Turbine Shaft Fatigue Life; Machine Design, Vol.4(2012) No.4, ISSN 1821-1259; pp. 231-236
[5] Peterson, R. E. (1959), Notch Sensitivity, Metal
Fatigue, Edited by G. Sines, J. L. Waismas, McGraw
Hill, pp. 293-306.
[6] Neuber, H. (1958), Theory of Notch Stresses,
translation of 1957 Edition in German, Springer,
Berlin.
[7] Suresh, S. (1998), Fatigue of Materials, Cambridge
University Press. ISBN 9780521578479, pp 30 -94
[8] El Haddad, M. H., Dowling, N.E., Topper, T. H., and
Smith. K. N. (1980), J Integral Applications for Sort
Fatigue Crack at Notches, International Journal of
Fracture, 16, 15.
[9] Taylor, D. (2007), Theory of Critical Distances: A
New Perspective in Fracture Mechanics, ISBN 978-008-044478-9, Elsevier Science
[10] Chattopadhyay, S, (2010), High Cycle Fatigue of
Structural Components Using Critical Distance
Methods, Proceedings of the SEM Annual
Conference June 7-10, 2010, ISBN 978-1-44199497-4, The Society for Experimental Mechanics,
Indianapolis, Indiana USA, pp 463 - 468
[11] Momčilović, D. et al. (2012). Failure analysis of
hydraulic turbine shaft. Engineering Failure Analysis,
Elsevier, Volume 20, March 2012, pp. 54-66.
[12] GOST 977-88, Steel Catings. General Specification
[13] Troschenko, VT. Fatigue resistance of metals and alloys.
Naukova Dumka; 1987. p. 668–70. [in Russian].
[14] Atanasovska, I., et al. (2010). FEM model for
calculation of Hydro turbine shaft. Proceedings of the
Sixth International Symposium KOD 2010, Palić,
Serbia, pp. 183-188.
236
[15] Atanasovska, I., et al.. (2011). Influence of transition
section of shaft with flange on stress concentration
factor. The 7th International scientific conference
Research and development of mechanical elements and
systems, April 2011., Zlatibor, Serbia, pp. 213-218.
[16] Bordeasu, I., Popoviciu, M. O., Novac, D. M., Baran,
M., (2009). Recorded cracks in the shaft of a
Transactions
on
hydraulic
bulb
turbine,
MECHANICS, SCIENTIFIC BULLETIN, University
of Timisoara, Romania, Tomul 54(68), Fascicola 2,
ISSN 1224-6077, pp. 19-25.
[17] Bordeasu, I., Popoviciu, M. O., Novac, D. M. (2009).
Fatigue studies upon horizontal hydraulic turbines
shafts and estimation of crack initiation, Journal of
Faculty of Technical Science, MACHINE DESIGN,
ISSN 1821-1259, pp183 – 186.
[18] Popoviciu, M. O., Bordeasy, I., (2011). Analyzes of
fissures in bulb turbine shafts, Journal of Faculty of
Technical
Science,
MACHINE
DESIGN,
Vol.3(2011) No.4, ISSN 1821-1259, pp. 233-240.
[19] Winston Revie R, Uhlig Herbert H. (2008) Corrosion
and corrosion control. John Wiley & Sons, Inc. pp.
173-174.
[20] Gangloff RP. (2005), Corrosion fatigue cracking, in
corrosion tests and standards: application and
interpretation. In: Baboian R, editor. ASM
International; [2nd ed.].p. 302–321
[21] Momčilović, D., Mitrović, R., Atanasovska, I. (2012).
Quantification of stress raiser influence of machine parts
fatigue life, Proceedings - 7th International Symposium
KOD 2012, 24-26 May 2012, Balatonfüred, Hungary,
ISBN 978-86-7892-399-9, pp. 215 – 218.